Will organisational and data silos soon become things of the past?

Silos are among the top three public enemies of the digital transformation narrative, alongside its accomplices, the lack of alignment with business strategy and human resistance to change. As data – captured in the past as well as in real time – is the lifeblood of digital technology, it stands to reason that ineffective, siloed flows of information are one of the major pitfalls that can hinder its implementation.

 

In business, silos are all over the place. Thanks to traditionally departmentalised and hierarchical organisational structures, the greatest divides are between different functions, which often focus on their own rather than overall company performance, and tend to have their own microcultures and set ways of how things must get done.

 

Needless to say, failure to share best practices or to see the department’s procedures as an integral part of the corporate big picture are bound to breed inefficiency and disjointed processes.

 

Agile frameworks originally developed for the fast-paced software development industry and then adopted more widely as a general project management methodology have already made great strides in rendering rigid departmental borderlines more porous since the 2000s.

 

In the agile methodology, teams are made up of employees from different functions that collectively have a diverse skillset up their sleeves. The different perspectives that members have of the organisation ensure that the team as a whole can transcend individual and departmental limitations and adopt a shared view of the organisation and its objectives.

 

The removal of data silos – both a prerequisite and an outcome

 

Digital data sitting in silos where it can’t be accessed is as much use as unwittingly owning a property without realising it. But the gradual migration of on-premises data centres into the cloud have had the added benefit of making businesses aware of the wealth of data they possess.

 

To prevent the migration exercise from becoming simple data dumping, companies have also had to cleanse their data to ensure its accuracy and relevance, as well as remove any duplicates and other redundancies.

 

If done well, data migration to the cloud can turn information sitting in silos into a goldmine of standardised and structured data that can be tapped into to gain actionable insights and improve decision making.

 

By 2023, 60 per cent of corporate data was already stored in the cloud, meaning more than half of synchronised and reliable global company data could be accessed by employees authorised to do so. And, as cloud data-storage services often also come with data analytics tools, a high percentage of this data can also be interrogated by the data owner company’s workforce to improve their processes and make data-driven decisions.

 

Access to the same data and the same consolidated system across the organisation fosters communication and dialogue between the different functional units of the business. A striking example of how this can improve a department’s operation is when access to the performance indicators of an upstream function can put a department’s output in a wider context.

 

A sales department with a sharp increase in its performance metric may learn, for example, that the positive trend reflects a new approach by the marketing team, rather than peak performance of their salespeople. Conversely, another department’s deteriorating performance may be traced back to another team’s mistake further up the value chain.

 

Gen AI’s role in consigning data silos to oblivion

 

The extraordinary data retrieval capabilities of large language models (LLMs) which provide the backbone of gen AI systems can turn enterprise data into easily searchable knowledge bases. If the enterprise database integrates data across different functions, anyone in any part of the organisation can have company data at their fingertips.

 

Thanks to gen AI’s language processing and understanding capabilities, employees can interrogate enterprise systems by asking them questions in conversational English – or any of the 50-plus languages gen AI understands – in the same way as they would a colleague conversant with the subject. 

 

Enterprise databases leverage another game-changing LLM feature – its capability to break down unstructured text, audio or video data into tokens and convert it into indexed data that can be analysed alongside with other structured data. This way not only structured enterprise data but also emails, presentations and meeting write-ups become readily searchable by anyone authorised.

 

As often is the case with digital technology, gen AI-assisted enterprise search also brings results while also reducing opportunities for human interaction. Technology enabling access to information, however, can also open new doors for interfunctional communication, collaborative projects and innovation that reach beyond team boundaries.

 

The demise of organisational and data silos doesn’t happen overnight, but in cycles. Sometimes an organisation adjusts its operation to the technology; at other times, new technological capabilities have a bearing on how businesses organise their workflows, creating a feedback loop that will eventually lead to the elimination of legacy system barriers. Often, the drawbacks of old practices are already well documented by the time the inner logic of a new technology turns change into an imperative – thus making old habits die sooner.